Parametric identification of distributed difference systems given observation interferences in input and output signals

Background. The article proposes an algorithm of parametric identification of a discrete stochastic system with distributed parameters in the presence of observation interferences in input and output signals in conditions of prior uncertainty (lack of information about observation interference distribution laws). It is well-known that the most routine method of estimation, the least square method, cannot be used as the stated problem doesn't refer to a regression class, and one cannot obtain consistent estimates of unknown parameters. Determination of consistent estimates for an ordinary stochastic difference equation with observation interferences was obtained by the author on the basis of the non-linear least-square method. The article gives an opportunity to apply independent estimation methods of ordinary difference equations for a discrete system with distributed constants. Matherials and methods. The article examines generalization of the non-linear least-square method for a discrete stochastic system with distributed parameters. Conclusions. This method can be used for parametric identification of stochastic difference equations with distributed parameters in the presence of observation interferences in input and output signals.